This address:n1JagbRWBDi6VMvG7HfZmXX74dB9eiHJzU (on testnet) has a negative balance of -0.27388239 according to: https://www.biteasy.com/testnet/addresses/n1JagbRWBDi6VMvG7HfZmXX74dB9eiHJzU, namely, it has sent more coins than it has received. Is that possible? Is it an error on the block explorer?
Short answer: Yes, this is probably a bug in the block explorer. blockexplorer.com shows a positive balance.
Some of the transactions shown there are "strange" (they presumably have non-standard scripts which someone was experimenting with), and this probably is tripping up the biteasy explorer.
Note that the notion of "balance" of an address only really makes sense when using "standard" transactions, where the output script says: to spend these coins, provide a public key whose hash is the address X, and a signature made by the corresponding private key. For fancier transactions, "balance" gets more fuzzy. For instance, I could make a transaction that requires signatures by two private keys, corresponding to addresses X and Y. Should those coins be "credited" to both addresses, or neither? Either way the books will not balance.
The negative balance was a bug which we have already fixed. You can now see that it's correct and agrees with the one from blockexplorer.com: https://www.biteasy.com/testnet/addresses/n1JagbRWBDi6VMvG7HfZmXX74dB9eiHJzU
Each wallet or
wallet.dat file consists of many separate accounts and each account can have many separate addresses. My understanding is that it is possible for a given account or a specific address to have a negative balance but not the wallet itself.
So for instance I could have a wallet with two accounts
bar and with balances of -4 btc and 5 btc respectively but if I query the balance of the wallet itself then it will show 1 btc. I could not however have an accont with 4 btc and the only other with -5 btc because this would leave the wallet with a negative balance.